search

Keck-DataReductionPipelines / KPF-Pipeline

KPF-Pipeline
https://kpf-pipeline.readthedocs.io/en/latest/
Other
11 stars 2 forks source link

Problems with QLP #786

Closed awhoward closed 8 months ago

awhoward commented 8 months ago

I figured out the problem with the QLP. There were three problems actually.

  1. L1 just gets behind (I think) or maybe crashes (I haven’t seen examples).
  2. When I tried to catch up L1, I didn’t use ncpu so my watch script wasn’t producing any output.
  3. Recent on-sky observations are missing one of the two Guider extensions recently, so no Guider QLP data products are made. I sent a message to Brad and Josh about this.

While figuring that out, I made the changes in the attached PR. In the end, these changes aren't that significant, but I think should be incorporated. They are 1) a few comments/example fixes, 2) ncpu is now set to 20 for L1 files and 10 for other files in the launch_QLP script, and 3) I renamed the QLP watch script from quicklook_watch_dir.recipe to quicklook_watch.recipe because it seems more consistent with the use of the word "watch" elsewhere.

awhoward commented 8 months ago

Here's the related issue: https://github.com/Keck-DataReductionPipelines/KPF-Pipeline/issues/782

howardisaacson commented 8 months ago

We have a page on Confluence titled Setting up the DRP and QLP. Does that page need updated with the changes that you made, @awhoward ?

awhoward commented 8 months ago

Good point, Howard. I checked and there's no correct information (related to this PR) on that page. I will add some updates to that page though related to catching up on QLP dataproducts.

awhoward commented 8 months ago

@bjfultn -- it would be good to merge this today or early tomorrow so that we can run the QLP from develop for observing tomorrow.